Predicting serious postoperative complications and evaluating racial fairness in machine learning algorithms for metabolic and bariatric surgery.

BACKGROUND: Predicting the risk of complications is critical in metabolic and bariatric surgery (MBS). OBJECTIVES: To develop machine learning (ML) models to predict serious postoperative complications of MBS and evaluate racial fairness of the models. SETTING: Metabolic and Bariatric Surgery Accreditation and Quality Improvement Program (MBSAQIP) national database, United States. METHODS: We developed logistic regression, random forest (RF), gradient-boosted tree (GBT), and XGBoost model using the MBSAQIP Participant Use Data File from 2016 to 2020. To address the class imbalance, we randomly undersampled the complication-negative class to match the complication-positive class. Model performance was evaluated using the area under the receiver operating characteristic curve (AUROC), precision, recall, and F1 score. Fairness across White and non-White patient groups was assessed using equal opportunity difference and disparate impact metrics. RESULTS: A total of 40,858 patients were included after undersampling the complication-negative class. The XGBoost model was the best-performing model in terms of AUROC; however, the difference was not statistically significant. While the F1 score and precision did not vary significantly across models, the RF exhibited better recall compared to the logistic regression. Surgery type was the most important feature to predict complications, followed by operative time. The logistic regression model had the best fairness metrics for race. CONCLUSIONS: The XGBoost model achieved the highest AUROC, albeit without a statistically significant difference. The RF may be useful when recall is the primary concern. Undersampling of the privileged group may improve the fairness of boosted tree models.

Exploring Doping Mechanisms and Modulating Carrier Concentration in Copper Iodide: Applications in Thermoelectric Materials.

Due to its small hole-effective mass, flexibility, and transparency, copper iodide (CuI) has emerged as a promising p-type alternative to the predominantly used n-type metal oxide semiconductors. However, the lack of effective doping methods hinders the utility of CuI in various applications. Sulfur (S)-doping through liquid iodination is previously reported to significantly enhance electrical conductivity up to 511 S cm-1. In this paper, the underlying doping mechanism with various S-dopants is explored, and suggested a method for controlling electrical conductivity, which is important to various applications, especially thermoelectric (TE) materials. Subsequently, electric and TE properties are systematically controlled by adjusting the carrier concentration from 3.0 × 1019 to 4.5 × 1020 cm-3, and accurately measured thermal conductivity with respect to carrier concentration and film thickness. Sulfur-doped CuI (CuI:S) thin films exhibited a maximum power factor of 5.76 µW cm-1 K-2 at a carrier concentration of 1.3 × 1020 cm-3, and a TE figure of merit (ZT) of 0.25. Furthermore, a transparent and flexible TE power generator is developed, with an impressive output power density of 43 nW cm-2 at a temperature differential of 30 K. Mechanical durability tests validated the potential of CuI:S films in transparent and flexible TE applications.

Behavioral biometric optical tactile sensor for instantaneous decoupling of dynamic touch signals in real time.

Decoupling dynamic touch signals in the optical tactile sensors is highly desired for behavioral tactile applications yet challenging because typical optical sensors mostly measure only static normal force and use imprecise multi-image averaging for dynamic force sensing. Here, we report a highly sensitive upconversion nanocrystals-based behavioral biometric optical tactile sensor that instantaneously and quantitatively decomposes dynamic touch signals into individual components of vertical normal and lateral shear force from a single image in real-time. By mimicking the sensory architecture of human skin, the unique luminescence signal obtained is axisymmetric for static normal forces and non-axisymmetric for dynamic shear forces. Our sensor demonstrates high spatio-temporal screening of small objects and recognizes fingerprints for authentication with high spatial-temporal resolution. Using a dynamic force discrimination machine learning framework, we realized a Braille-to-Speech translation system and a next-generation dynamic biometric recognition system for handwriting.

Junctionless Structure Indium-Tin Oxide Thin-Film Transistors Enabling Enhanced Mechanical and Contact Stability.

In recent years, considerable attention has focused on high-performance and flexible crystalline metal oxide thin-film transistors (TFTs). However, achieving both high performance and flexibility in semiconductor devices is challenging due to the inherently conductive and brittle nature of crystalline metal oxide. In this study, we propose a facile way to overcome this limitation by employing a junctionless (JL) TFT structure via oxygen plasma treatment of the crystalline indium-tin oxide (ITO) films. The oxygen plasma treatment significantly reduced oxygen vacancies in the ITO films, contributing to the significant reduction in the carrier concentration from 4.67 × 1020 to 1.39 × 1016. Importantly, this reduction was achieved without inducing any noticeable structural changes in the ITO, enabling the successful realization of ITO JL TFTs with an adjustable threshold voltage. Furthermore, the ITO JL TFTs demonstrate good stability and reliability under various bias stress conditions, aging in the air atmosphere, and high-temperature processes. In addition, the ITO JL TFTs exhibit low light sensitivity due to the wide bandgap of ITO and further suppression of Vo defects, making them suitable for applications requiring stable performance under light exposure. To compare and analyze the flexibility of the JL structure and conventional structure with additional source/drain (S/D) junction in ITO TFTs with nonencapsulation, we utilized mechanical simulations and transmission line method (TLM). By employing the JL structure in ITO TFT through carefully optimized oxygen plasma treatment, we successfully mitigated stress concentration at the S/D-channel interface. This resulted in a JL ITO TFT that exhibited a change in contact resistance of less than 20% even after 20,000 bending cycles. Consequently, a stable and flexible ITO TFT with field-effect mobility (µFE) of 12.74 cm2/(V s) was realized, outperforming conventionally structured ITO TFTs with additional S/D junction, where the contact resistance nearly tripled.

Cost-effectiveness analysis of the 20-valent pneumococcal conjugate vaccine for the pediatric population in South Korea.

OBJECTIVES: To evaluate the cost-effectiveness of the 20-valent pneumococcal conjugate vaccine (PCV20) compared to 13-valent pneumococcal conjugate vaccine (PCV13) for the pediatric population in Korea, where the four-dose vaccine coverage rate is over 97%. METHODS: We constructed a Markov model to calculate the cost and quality-adjusted life-years (QALYs) over 10 years. The health states were susceptible states; disease states, which included invasive pneumococcal diseases such as meningitis, bacteremia, pneumonia, and acute otitis media; and death attributable to pneumococcal disease. The annual incidence and mortality due to pneumococcal diseases were estimated based on the serotypes covered by PCV13 and PCV20, vaccine coverage rate, vaccine effectiveness, and population size. Vaccine, administration, and disease costs were included in the model. RESULTS: In the total population (n = 51,431,305), PCV20 prevented more pneumococcal diseases and deaths, resulting in a gain of 74,855 QALY over PCV13. Meanwhile, the PCV20 group spent $275,136,631 less than the PCV13 group. As PCV20 gained more QALYs but spent less on total medical costs than PCV13, PCV20 was dominant over PCV13. CONCLUSIONS: In the Korean population, PCV20 is a cost-effective and dominant option over PCV13. Our findings provide evidence for decision-making regarding the introduction of PCV20 in countries with high vaccine coverage.

Photoexcited Carrier Dynamics in Iodine-Doped CH3NH3PbBr3 Single Crystals.

Iodine-doped bromide perovskite single crystals (IBPSCs) have important applications in optoelectronic fields, such as in solar cells. Currently, much research has aimed to study the phase separation phenomenon and device performance improvements in IBPSCs. However, important intrinsic photoexcited carrier dynamics are often overlooked in IBPSCs. Here, we explored the photoexcited carrier dynamics in typical iodine-doped MAPbBr3 single crystals using the excitation intensity-dependent steady-state photoluminescence (PL) and time-resolved photoluminescence (TRPL) technique. We found that the trap state density changes with an increase in the amount of doped iodine. Further, we noticed that there is an influence of carrier diffusion on the photoexcited carrier dynamics, and then, we evaluated the carrier diffusion coefficients and recombination constants via numerical simulations of the PL kinetics. Consequently, we found that the electron shallow trap-related carrier behaviors substantially impacted the PL kinetics. Our results greatly facilitate a deeper understanding of the fundamental characteristics of mixed halide perovskite material.

Disturbed flow regulates protein disulfide isomerase A1 expression via microRNA-204.

Redox processes can modulate vascular pathophysiology. The endoplasmic reticulum redox chaperone protein disulfide isomerase A1 (PDIA1) is overexpressed during vascular proliferative diseases, regulating thrombus formation, endoplasmic reticulum stress adaptation, and structural remodeling. However, both protective and deleterious vascular effects have been reported for PDIA1, depending on the cell type and underlying vascular condition. Further understanding of this question is hampered by the poorly studied mechanisms underlying PDIA1 expression regulation. Here, we showed that PDIA1 mRNA and protein levels were upregulated (average 5-fold) in the intima and media/adventitia following partial carotid ligation (PCL). Our search identified that miR-204-5p and miR-211-5p (miR-204/211), two broadly conserved miRNAs, share PDIA1 as a potential target. MiR-204/211 was downregulated in vascular layers following PCL. In isolated endothelial cells, gain-of-function experiments of miR-204 with miR mimic decreased PDIA1 mRNA while having negligible effects on markers of endothelial activation/stress response. Similar effects were observed in vascular smooth muscle cells (VSMCs). Furthermore, PDIA1 downregulation by miR-204 decreased levels of the VSMC contractile differentiation markers. In addition, PDIA1 overexpression prevented VSMC dedifferentiation by miR-204. Collectively, we report a new mechanism for PDIA1 regulation through miR-204 and identify its relevance in a model of vascular disease playing a role in VSMC differentiation. This mechanism may be regulated in distinct stages of atherosclerosis and provide a potential therapeutic target.

A nationwide analysis of the treatment patterns, survival, and medical costs in Korean patients with relapsed or refractory diffuse large B-cell lymphoma.

Background: Approximately one-third of patients with diffuse large B-cell lymphoma (DLBCL) are refractory to treatment or experience relapse after initial therapy. Unfortunately, treatment options for older patients and those who experience relapse or become refractory to hematopoietic stem cell transplantation (HSCT) are limited. This nationwide population-based study aimed to identify treatment patterns, survival times, and treatment costs in patients with relapsed/refractory DLBCL (R/R DLBCL). Materials and methods: Between 2011 and 2020, data on patients with R/R DLBCL were retrieved from the Korean Health Insurance Review & Assessment Service, encompassing the entire population. We identified the treatment patterns for each treatment line using a Sankey diagram and calculated the median time to the subsequent treatment in line. Median overall and progression-free survival times were estimated using the Kaplan-Meier survival curves. Finally, the medical costs incurred during DLBCL treatment were calculated for each treatment line and the costs related to HSCT were summarized at the episode level. Results: A total of 864 patients with R/R DLBCL who received second-line treatment were identified, and a regimen of ifosfamide, carboplatin, and etoposide (ICE) was administered the most. Among them, 353 were refractory or relapsed cases that were treated with third-line treatments. The median times for second-line to third-line, third-line to fourth-line, fourth-line to fifth-line, and fifth-line to sixth-line treatment failures gradually decreased (3.93, 2.86, 1.81, and 1.38 months, respectively). The median overall survival time was 8.90 and 4.73 months following the second-line and third-line treatments, respectively. In the third-line treatment setting, the patients did not show a significant difference in survival time after HSCT. The median medical cost was $39,491 across all treatment lines including the cost of HSCT which was $22,054. Conclusion: The treatment patterns in patients with R/R DLBCL, especially at third-line treatments and thereafter, were complicated, and their prognosis was poor despite the high medical costs. Novel and effective treatment options are expected to improve the prognosis and alleviate the economic burden of patients with R/R DLBCL.

Predicting operative time for metabolic and bariatric surgery using machine learning models: a retrospective observational study.

BACKGROUND: Predicting operative time is essential for scheduling surgery and managing the operating room. This study aimed to develop machine learning (ML) models to predict the operative time for metabolic and bariatric surgery (MBS) and to compare each model. METHODS: The authors used the Metabolic and Bariatric Surgery Accreditation and Quality Improvement Program database between 2016 and 2020 to develop ML models, including linear regression, random forest, support vector machine, gradient-boosted tree, and XGBoost model. Patient characteristics and surgical features were included as variables in the model. The authors used the mean absolute error, root mean square error, and R 2 score to evaluate model performance. The authors identified the 10 most important variables in the best-performing model using the Shapley Additive exPlanations algorithm. RESULTS: In total, 668 723 patients were included in the study. The XGBoost model outperformed the other ML models, with the lowest root mean square error and highest R 2 score. Random forest performed better than linear regression. The relative performance of the ML algorithms remained consistent across the models, regardless of the surgery type. The surgery type and surgical approach were the most important features to predict the operative time; specifically, sleeve gastrectomy (vs. Roux-en-Y gastric bypass) and the laparoscopic approach (vs. robotic-assisted approach) were associated with a shorter operative time. CONCLUSIONS: The XGBoost model best predicted the operative time for MBS among the ML models examined. Our findings can be useful in managing the operating room scheduling and in developing software tools to predict the operative times of MBS in clinical settings.

Cost-effectiveness of the 20-valent pneumococcal conjugate vaccine versus the 23-valent pneumococcal polysaccharide vaccine for older adults in South Korea.

BACKGROUND: Despite the demonstrated immunogenicity and safety of the 20-valent pneumococcal conjugate vaccine (PCV20) in older adults, the cost-effectiveness of the PCV20 was not examined compared to the 23-valent pneumococcal polysaccharide vaccine (PPSV23) in South Korea. Therefore, this study aimed to evaluate the cost-effectiveness of PCV20 compared with PPSV23 in adults aged 65 years and older in South Korea. METHODS: We constructed a Markov model that included susceptible states, invasive pneumococcal disease (IPD), non-bacteremic pneumonia (NBP), and death. The population was categorized by disease risk status (low risk, moderate risk, and high risk) and age group (65-74/75-84/85-99 years) at model entry. The annual incidence and mortality of IPD and NBP associated with PCV20 and PPSV23 were estimated based on serotype coverage, vaccine coverage, and vaccine effectiveness. The disease costs and utilities were obtained from previous studies. The incremental cost-effectiveness ratio (ICER) was used to evaluate cost-effectiveness within the threshold of 16,824 USD per quality-adjusted life-year (QALY). RESULTS: Among the total population (n = 8,843,072), PCV20 prevented 1941 and 50,575 cases of IPDs and NBPs, respectively, and 898 and 8593 deaths due to IPDs and NBPs compared to PPSV23. The total medical cost per person was 12.11 USD higher in PCV20, with a gain of 0.0053 LYs and 0.0045 QALYs per person. The ICER for PCV20 and PPSV23 was 2270 USD/LY and 2677 USD/QALY. CONCLUSIONS: In South Korea, PCV20 is a cost-effective option compared with PPSV23 for adults aged 65 years and older. These cost-effectiveness results provide evidence for decision-making regarding the approval and National Immunization Program implementation of PCV20.